Skill UI

A practical decision framework for determining whether to use regular expressions or Large Language Models (LLMs) when parsing structured text (e.g., forms, quizzes, invoices). It recommends starting with deterministic, low-cost regex for 95-98% of structured data, and only invoking expensive LLM calls for identifying and validating ambiguous edge cases, optimizing both cost and accuracy.

This skill detects covert data exfiltration occurring over DNS tunnels. It analyzes DNS query names by computing Shannon entropy, examining query length distributions, and identifying high subdomain cardinality. It is essential for security operations, incident response, and proactive threat hunting when suspicious or encoded DNS activity is suspected.

A practical decision framework for determining whether to use regular expressions or Large Language Models (LLMs) when parsing structured text (e.g., forms, quizzes, invoices). It recommends starting with deterministic, low-cost regex for 95-98% of structured data, and only invoking expensive LLM calls for identifying and validating ambiguous edge cases, optimizing both cost and accuracy.

A comprehensive guide for authors submitting papers related to web search, ranking, and data mining to WSDM. It provides critical advice on conference fit assessment, proper manuscript re-framing, required evidence (retrieval/ranking baselines, platform-aware evaluation), and how to position the work against competing top venues like SIGIR and KDD.

A comprehensive guide for authors submitting data mining papers to the ACM SIGKDD Conference. It helps authors assess the manuscript's fit, reframe the contribution (e.g., from journal to conference style), and strengthens the evidence bar by focusing on novelty, scale, reproducibility, and clear real-world impact required by top-tier data science venues.

A comprehensive guide for designing rigorous and robust empirical programs in Information Retrieval, Data Mining, and Knowledge Management. It outlines how to meet the stringent evidence requirements for top-tier conferences, ensuring claims are substantiated across multiple evaluation dimensions (retrieval, mechanism, scalability) and incorporating best practices for reproducibility and handling modern challenges like LLM contamination.

This guide details the complex requirements for writing the 'Related Work' section for CIKM submissions. It emphasizes the need for 'three-front coverage' (Retrieval, Mining, and Knowledge Management/Databases) and provides specific instructions on writing the 'boundary-work paragraph.' It also covers critical academic hygiene, such as avoiding venue misattribution, maintaining recency baselines, and accurately stating novel contributions.

This guide helps researchers assess if their work correctly spans the tri-community area of Information Retrieval, Data Mining, and Knowledge Management. It provides a detailed framework for comparing CIKM against major academic venues like SIGIR, KDD, WSDM, and SIGMOD/VLDB, ensuring the paper's scope is optimally matched to the target conference.

This skill provides a strategic guide for authors evaluating if their AI/data mining manuscript is a suitable fit for the ECML PKDD conference. It covers scope definition, evidence requirements, academic framing, and strategic positioning relative to major global venues (e.g., ICML, NeurIPS). Use this to refine your submission narrative and maximize your chances of acceptance.

This skill guides authors in determining the suitability of their manuscripts for the journal Genome Biology. It details the journal's scope, which requires integrating large-scale genomic or epigenomic data with novel computational methods or statistical frameworks. It emphasizes that successful submissions must provide both methodological advances and meaningful biological insights, adhering strictly to standards like open-source code, public data deposition (GEO, SRA, ENCODE), and rigorous benchmarking.

Provides comprehensive guidelines for packaging research artifacts (code, data, logs) for submission to data mining conferences like ICDM. The focus is on ensuring the artifact is completely anonymous for triple-blind review (scrubbing Git history, paths, and internal names) and maximizing reproducibility, allowing reviewers to verify results with minimal effort.

A comprehensive guide for authors preparing submissions for top-tier data mining conferences (e.g., ICDM). It emphasizes the critical strategy of preemptively addressing potential weaknesses—such as baselines, variance, scaling, and data leakage—directly within the submitted manuscript, as author rebuttals are often not permitted. Also provides strict guidelines for drafting anonymous, high-impact responses if a rebuttal window is available.